digital-alloy-lab

Digital alloy development at the Vertical Cloud Lab @ BYU: AI-guided design of aluminum alloys for laser powder bed fusion (FLAIME), Bayesian optimization over function networks (BOFN), CALPHAD, powder atomization, and the proposals built on them.

This repo was split out of the private vertical-cloud-lab/digital-alloy-lab-private in October 2026.

• Issues were transferred here and kept their numbers, so old issue links redirect.
• Open pull requests were recreated here from the same branches, with the same numbers. Their review comments and commit-by-commit history stay in the private repo.
• Closed pull requests have a closed placeholder issue at their number, so every #N still lands on the right item.
• Files that stay private were left out: signed letters, per-person budget detail, CVs and similar. Publisher PDFs are listed by DOI in literature/DOIs.md instead.

Contents

• Alloy Workflow Review — Comprehensive review of the alloy synthesis, processing, characterization, and performance pipeline, including current status, equipment needs, and funding strategy. References issues and PRs in vertical-cloud-lab/byu-vcl.
• Grant Targeting Analysis — Analysis of which grant program to target initially (NSF CAREER, DARPA DSO BAA, NSF DMR TMRP, etc.), informed by Edison Scientific literature research on program managers, funded programs, competitive landscape, key reviewers, publication venues, simulation tools, and prior literature.
• Deep Literature Search Synthesis — Synthesis of eight Edison Scientific LITERATURE_HIGH queries covering self-driving labs, powder production, automated characterization, CALPHAD for AM, ML for LPBF process-structure-property, digital twins, function networks (Frazier), and LLM-enhanced Bayesian optimization. Includes gap analysis and recommended next steps for the autonomous AM alloy discovery pipeline.
• BOFN + LLM-BO Synthesis — Synthesis of Bayesian Optimization over Function Networks (BOFN) and LLM-augmented BO, including a verified BoTorch implementation, 5 proposed integration points for LLM-BOFN, and a 4-tier benchmark task design for building confidence before deployment.
• Company Landscape — Identification and categorization of companies in autonomous alloy discovery, high-throughput AM, and AI-accelerated materials development, including Genesis Mission partnership potential for Topic 3B (Critical Minerals, AMMTO).
• High-Throughput CALPHAD — Assessment of CALPHAD/Thermo-Calc high-throughput capabilities on HPC clusters, covering open-source (pycalphad, Equilipy) and commercial (TC-Python) tools with parallelization strategies.
• Thermo-Calc Discussion Transcript — Meeting transcript (April 9, 2026) between Thermo-Calc's account manager, a Thermo-Calc specialist, and Sterling Baird regarding Genesis Mission and Thermo-Calc collaboration.
• Alloy Databases — Survey of databases for alloy properties, compositions, and materials criticality relevant to AI-guided alloy design campaigns targeting DOE Genesis Mission Topic 3B (Critical Minerals, AMMTO). Covers aluminum, magnesium, and other alloy families plus critical mineral supply risk databases (USGS, DOE CMS, EU JRC, alloy-level criticality indices). Includes access model tracking (open, proprietary, educational), vendor contacts, and integration points with FLAIME's pKGFN pipeline and Thermo-Calc collaboration.
• Database Samples — Small data subsets downloaded from accessible databases (AFLOW, OQMD, NOMAD, JARVIS, Figshare, Harvard Dataverse, Zenodo, pycalphad, ULTERA) to verify access and illustrate data formats.
• Benchmarks — Working implementations of optimization methods, including a 4-node BOFN demo (bofn_demo.py) verified with BoTorch 0.17.2.
• Figures — Publication-quality figure generation scripts and outputs, including LPBF parameter space visualization and panel-based composites.
• Genesis Mission Gotchas — Analysis of factors that reduce likelihood of FLAIME proposal selection, covering PI profile, institutional factors, preliminary data, team composition, competitive landscape, and program-specific risks for Topic 3B.
• Genesis Mission Topic 12 / Focus Area F — (Superseded — retained for historical reference) Original analysis when FLAIME targeted Topic 12F, prior to the Topic 3B pivot.
• Genesis Mission Proposal — LaTeX proposal template, compiled PDFs, Edison template review, and proposal checklist for DOE DE-FOA-0003612 Topic 3B submission.
• TMS 2027 Ultrasonic Atomization Abstract — Conference abstract for TMS 2027 on tuning and modeling ultrasonic atomization parameters (vibration amplitude, melt superheat, atmosphere) for critical-mineral-lean aluminum LPBF powders on the AMAZEMET rePOWDER, leaning into a multi-run, multi-alloy dataset (commercial + novel lean compositions) feeding closed-loop AI optimization. Includes a submission-ready ~150-word version, an expanded alternate, a recommended-symposium analysis (primary: Atomization & Powder Metallurgy for Advanced Applications — Honoring Iver Anderson), and a pre-submission checklist.
• Publication Strategy — Prioritized publication roadmap (16 papers, including 2 new 3B-native papers) for maximizing proposal fundability under DOE Genesis Mission Topic 3B (Critical Minerals, AMMTO). Partnership-integrated across 3 Letters of Collaboration (Valimet, AMAZEMET, Thermo-Calc) and 8 Letters of Support (Ames, Fehrmann, Sparks, Speer, Clare, Stiles, Hattrick-Simpers, Brown). Includes mineral-burden-aware BOFN, critical-mineral-lean Al alloy perspectives, mock PO/reviewer feedback, incremental sequencing, ROI analysis, grant-specific mapping, Partner × Publication contribution matrix, and preliminary results integration (atomizer installed, AlSi10Mg tensile from U of Utah/3D Powder Tech, SEM/EDS, pKGFN pipeline, Thermo-Calc AM Module sims, automated powder dispensing). Informed by two Edison Scientific LITERATURE_HIGH queries (48K + 48K chars, 57 combined citations) and two Edison ANALYSIS reviews, plus full repository context including the Topic 3B deep-dive series (PR #26).
• TMS 2027 LPBF Process-Parameter Optimization Abstract — TMS 2027 conference abstract for the FLAIME LPBF process-parameter-qualification work: a Thermo-Calc-AM-Module-seeded, Bayesian-optimization loop validated on an Aconity MIDI at the University of Utah (single tracks → density cubes → ISO/ASTM artifacts → tensile bars, per powder), with an explicit cross-machine/cross-instance transferability study and an LLM-agent pilot. Includes a 147-word submission version, expanded and referenced versions, a symposium-fit analysis (selected: AI-Enabled Materials Processing), a pre-submission checklist, and mock organizer review. Companion (downstream, print-side) to the ultrasonic atomization abstract (PR #98). Literature scoping and reviews under edison-trajectories/tms-2027-lpbf-parameter-abstract-review/, including the full-depth Edison Scientific LITERATURE_HIGH report (edison-literature-high-report.md, ~54 K chars, 20 ranked references).
• Edison Trajectories — Full Edison Scientific (PaperQA3 LITERATURE_HIGH) research outputs covering program managers, funded programs, best practices, reviewers, venues, alloy discovery literature, LPBF literature, simulation methods, self-driving labs, powder production, automated characterization, CALPHAD for AM, ML for LPBF PSP, digital twins for AM, function networks for BO, LLM-enhanced Bayesian optimization, alloy database surveys, materials criticality databases, Thermo-Calc AM module assessment, and publication strategy.
  • Thermo-Calc AM Module — Literature review and fit assessment for Thermo-Calc's Additive Manufacturing Module, covering published research (2020–2026), capabilities, limitations, and VCL integration recommendations.